Lymphocyte numbers are regulated both by responses to conventional exogenous antigens and by stimulation by endogenous peptide/MHC complexes. This joint regulation allows individuals to maintain a broad repertoire of specificities, allowing responses against a vast array of foreign entities and, at the same time, providing a pattern of memory based on the immunization history of the individual. The study of the process of lymphocyte dynamics that underlies this regulation requires a multidisciplinary approach, aimed both at the molecular underpinnings of the process through which lymphocytes survive and proliferate and a systemics approach to appreciate the overall mechanisms governing total numbers of lymphocytes of distinct phenotype and distinct specificity. Emphasis has been placed on three aspects of this problem: the dynamics of lymphocyte memory, the mechanisms underlying CD4 T cell depletion in HIV infection and the process of homeostatic proliferation. During the past year, it has been shown that although naive and memory cells of a given specificity both proliferate vigorously when challenged with their cognate antigens in vivo, naive cells accumulate to much larger numbers than memory cells. This process appears to be linked to the observation that memory cells may actually be depleted in number upon challenge, after they have undergone a rapid response that includes transient cytokine production and proliferation. This has led to the postulate that CD4 T cells are organized in a hierarchical system and that memory cells exist in a series of distinct states in which the most "immature" has the greatest renewal capacity and the most mature, the greatest death rate. This produces a population of T cells of a given antigenic specificity that, while capable of self-renewal, is also continually being replaced by cells moving from more immature pools, including the pool of naive cells. This has also led to an analysis of the mechanisms underlying CD4 T cell depletion in infections with HIV and with other agents that establish a chronic presence. One reason for such a depletion may be the development of a chronic immune/inflammatory response that results in resetting the stable set points of CD4 numbers so that their "depletion" is a result of the response to the agent and its inflammation-inducing properties and not necessarily due to infection of the CD4 cells themselves. Homeostatic proliferation has been examined in detail in neonatal settings, where it is clearly physiologic. It has been shown that this proliferation results in a population of naive lymphocytes developing into memory CD4 T cells. The proliferation of these cells depends upon interaction with endogenous peptide/MHC complexes. It is strongly inhibited by the presence of memory cells that recognize the same set of peptide/MHC complexes so that it results in the development of memory cell populations of very broad specificity. Indeed, it can be argued that homeostatic proliferation evolved not so much to regulate overall lymphocyte numbers but rather to ensure that an individual had a memory cell population of a broad and potentially useful repertoire.